Nonpolyhedral proof of the Michael finite-dimensional selection theorem (Q950734)

In the 1950's, \textit{E. Michael} proved a series of selection theorems, which became a powerful instrument in topology: the 0-dimensional selection theorem [Ann. Math. (2) 63, 361-382 (1956; Zbl 0071.15902)] in general topology; the convex selection theorem [ibid. 64, 562--580 (1956; Zbl 0073.17702)] in infinite-dimensional topology and analysis, and the finite-dimensional selection theorem [ibid. 65, 375--390 (1957; Zbl 0088.15003)] in geometric topology. The first two theorems have a short and simple proofs, however the proof of the third one is quite difficult. Trying to simplify the proof of the finite-dimensional selection theorem, the author of the paper under review gives a different way (still long) to prove it. His new approach is a nonpolyhedral proof (i.e., free of using the passage to nerves of covers). The author also gives an outline of a generalization of the filtered finite-dimensional selection theorem proved by \textit{E. V. Shchepin}, and \textit{N. B. Brodskij} [Proc. Steklov Inst. Math. 212, 209--229 (1996); translation from Tr. Mat. Inst. Steklova 212, 220--240 (1996; Zbl 0873.54022)] and gives a new proof of a theorem by \textit{V. V. Uspenskii} [Topology Appl. 85, No.~1--3, 351--374 (1998; Zbl 0921.54014)] on the existence of selections of maps with contractible images.